Improving Skills Development and Employability through Traditional Apprenticeships: Evidence from Senegal

The PEJA project (“Improving Youth Employability through Informal Apprenticeships”) seeks to facilitate skills transfer to apprentices by providing equipment grants to master craftspersons, along with providing additional training (technical skills, business skills and socio-emotional skills) to craftspersons and apprentices alike, and then introducing additional incentives for apprentices to complete formal certification of their skills. The main beneficiaries are young informal apprentices, aged 15 to 25, and their masters. It is worth noting that the program does not place youth in workshops: beneficiaries are apprentices who have selected themselves into apprenticeships before the beginning of the program. The project targets urban areas, where traditional apprenticeships are more prevalent: it is implemented in the four departments of the Dakar region and in 15 regional capitals or economic hubs (Diourbel, Fatick, Kaffrine, Kaolack, Kedougou, Kolda, Louga, Matam, Mbacke, Mbour, Saint-Louis, Sedhiou, Tambacounda, Thies, and Ziguinchor). In its first wave the program targeted workshops from 8 trades (tailoring, metalwork, woodwork, auto-mechanics, hairdressing, market gardening, refrigeration/air conditioning, processing of fruits and vegetables).

At the level of the apprentice, the PEJA consists of the following interventions:
- Literacy training for illiterate apprentices. Apprentices were assigned to this treatment based on their level of formal education.
- Business skills training. Business skills training was provided in class. It followed the Start and Improve your Business curricula (with a focus on “Generate your business idea” and “Start your business”) adapted for the apprentice population.
- Socio-emotional skills training. The training was provided in class and covered self-esteem, self-affirmation, sense of initiative, compliance with commitment, risk-taking, communication, problem solving, perseverance, and networking.
- Two one-time cash transfers to enable youth to embark on a path of self-employment. First, an economic inclusion cash transfer (US$200, with a 50 percent premium for women) provided to young apprentices, conditional on participation in the business training. Second, a performance-based cash transfer (US$216 with a 100 percent premium for women) provided to apprentices who have obtained validation of their skills.

At the master craftsperson level, the PEJA interventions aimed at benefiting apprentices indirectly through an improvement of their learning environment. This includes the following components:
- Technical training for masters. Once the masters’ needs have been identified, the project supported a technical training (5 days in class). This activity was expected to expand the scope of technical skills transferred and to improve the quality of apprenticeship training.
- Business skills training. Business skills training was provided in class. It followed the Start and Improve your Business curricula (with a focus on “Improve your business”).
- Socio-emotional skills training. The training was provided in class and covered self-esteem, self-affirmation, sense of initiative, compliance with commitment, risk-taking, communication, problem solving, stress management, teamwork, conflict resolution and leadership.
- Pedagogical skills training. This training aimed at improving the transmission of technical skills to apprentices.
- Capital grant to upgrade the technology used in the workshop. Equipment (worth US$2,000 on average) was provided to masters after an assessment of their needs during the business skills training.

2021-09-01

2024-03-31

At the apprentice level, the main outcomes of interest will be:
- Test scores: Mathematics and reading test scores, Socio-Emotional Skills scores. (decision making, personal initiative, perseverance, self-control, expressiveness, collaboration, negotiation. and listening), technical skills. business skills.
- Status with regard to apprenticeships training: Apprentice still in training as of endline survey, Total training duration as of endline survey, Successful completion of the apprenticeships training, Apprentice was certified since baseline.
- Employment and income: Employment status as of endline survey, work intensity, type of employment (self-employment, employee, family worker), sector of employment, income from main and secondary occupation, Entrepreneurship (apprentice has created his/her own business as of endline survey).
- Exposure to physical discipline.

At the master/ workshop level
- Tests scores: Business skills scores, Socio-emotional skills.
- Status in the profession: Master artisan is still operating his business as of the endline survey, Number of apprentices in charge as of endline survey, Master artisans got certified since baseline.
- Attitudes toward disciplinary measures
- Business performance: Business failure as of endline survey, Logarithm of sales, profit and master’s income.
- Business structure / vertical integration: Business size, Business workforce composition, Retention of apprentices, Human capital of new apprentices.

(detailed in the pre-analysis plan)

At the apprentice level, we will measure the following secondary outcomes:
- Marital status
- Motherhood and fatherhood
- Migration out of Senegal at the time of the survey.

This study follows a cohort 6129 apprentices and 2144 master-craftpersons over time, with a baseline survey (December 2020-January 2021) and a endline survey (September 2023 – March 2024). The identification strategy for the impact of the PEJA relies on the random assignment of masters and apprentices, eligible to the PEJA, to one of the two treatment groups or to the control group. The unit of selection is the workshop: when a workshop is selected a master and his/her apprentices are all assigned to the same group. The two treatment arms are (1) a full package and (2) an alternative package that includes all the components of the full package but the socio-emotional training for masters and apprentices.

At both rounds the survey covered a whole range of apprentices and masters characteristics, including mathematics and reading skills (for apprentices only), socio-emotional skills, business skills, technical skills (for apprentices only), and attitude regarding disciplining measures.

We will estimate the effect of the PEJA by comparing learning and employment outcomes between apprentices assigned to the to the full package treatment group and the control group. The specific impact of socio-emotional training will be obtained by comparing apprentices and masters of the two treatment groups.

Additionally, to estimate the effects of any spillovers and/or displacement effects, we will compare outcomes of apprentices in control neighborhoods to those from control workshops within treatment neighborhoods. Indeed, for tailors only, the assignment to treatment was done through a two-step sampling process, with selection clusters first randomly assigned to treatment and control and then treated workshops randomly selected in treatment selection clusters only.

Not available

Assignment to the two treatment groups was donein February 2021 using Stata and QGIS. The selection protocol differs for tailors and the other four trades. The selection was performed at the workshop level. When a workshop is selected, the master and his/her apprentices are all beneficiaries of the same treatment arm. In the few cases where two sampled masters were in the same workshop, the two of them (and their respective apprentices) are allocated to the same group.

For woodworkers, metalworkers, auto-mechanics and hairdressers, the selection was stratified by trade and localities (19 localities).

For tailors, the selection followed a more complex protocol to assure variation in the distance to the nearest treated workshop:
1. First, we identified proximity clusters using workshop GPS locations. Two workshops within 300 meters of each other (240 meters in secondary cities) are considered part of the same cluster.
2. The 284 identified clusters were classified into three categories depending on their size: 157 clusters with only one workshop, 115 middle size clusters and 12 large clusters.
3. For each category, a sixth of the clusters were randomly selected to be pure controls, meaning that all workshops were allocated to the control group.
4. In the remaining clusters, in each locality, workshops were randomly assigned to the two treatment groups.

The selection was done at the workshop level. When a workshop is selected, the master and his/her apprentices are all beneficiaries of the same treatment arm. In the few cases where two sampled masters were in the same workshop, the two of them (and their respective apprentices) are allocated to the same group.

Yes

2124 workshops. 1 to 4 apprentices were interviewed per workshop.

At baseline: 6129 apprentices, under the supervision of 2144 master craftspersons, in 2124 workshops.

- 2166 apprentices and 754 masters (in 745 workshops) in the full package treatment
- 1990 apprentices and 696 masters (in 690 workshops) in the alternative treatment group
- 1973 apprentices and 694 masters (in 689 workshops) in the control group

Under the assumption that 15% of apprentices may not be found by the end-line survey, the minimum detectable effect on probability of employment ranges between 4.0 and 5.8 percentage points (or 0.10 to 0.14 standard deviations) depending on the level of intra-cluster correlation within workshop clusters. (more details in the pre-analysis plan)

Pre-Analysis Plan